Reivindicaciones
CLAIMS What is claimed is: 1. A computational method for identifying one or more existing or novel chemical synthesis routes for producing a target compound comprising: determining a plurality of known chemical reactions and/or a plurality of novel chemical reactions, wherein the plurality of novel reactions are extrapolated from generalized known chemical transformations; determining, from the plurality of novel chemical reactions, a plurality of predicted chemical reactions, based on a trained classifier, wherein the trained classifier is trained on data derived from a plurality of chemical reactions known to be successful and a plurality of chemical reactions known to be unsuccessful that are instances of a given chemical transformation; generating a plurality of chemical reactions, based on the plurality of predicted chemical reactions and the plurality of known chemical reactions, wherein each chemical transition of the plurality of chemical reactions represents a transformation of one compound to another compound; determining at least one target compound; determining a plurality of chemical reaction routes associated with the at least one target compound, wherein each chemical reaction route comprises one or more chemical reactions of the plurality of chemical reactions that produces the target compound; and determining one or more optimal chemical reaction routes from the plurality of chemical reaction routes identified for producing the target compound wherein at least one of the one or more optimal chemical reaction routes comprises at least one known reaction transformation and at least one predicted reaction transformation. 2. The method of claim 1, further comprising training a classifier on a training data set, wherein the training data set comprises one or more of, a chemical reaction database, estimated yields, or predicted yields for the one or more chemical reactions. 3. The method of claim 2, wherein training the classifier on the training data set comprises: receiving a dataset comprising one or more chemical reactions based on one or more chemical transformations, wherein each of the one or more chemical reactions comprises at least one reactant, wherein each reactant is comprised of one or more atoms; for each reactant, classifying the one or more atoms into a category based on a neighborhood atom, a bond order, and/or a number of hydrogen atoms present; for each reactant, determining a vector based on a histogram of categories; determining a training dataset comprised of a) vectors of reactions associated with a specific transformation and b) vectors of reactions associated with the specific transformation but yield a product from a different reaction type; exposing a classifier to a portion of the training dataset to train the classifier; and exposing the trained classifier to another portion of the training dataset to test the trained classifier. 4. The method of claim 3, wherein exposing the trained classifier to another portion of the training dataset to test the trained classifier comprises assessing performance of the trained classifier based on one or more metrics. 5. The method of claim 4, wherein the one or more metrics comprise one or more of accuracy, positive precision, negative precision, positive recall, or negative recall. 6. The method of claim 1, further comprising generating a tree data structure, wherein the target compound is a root node of the tree data structure. 7. The method of claim 6, further comprising adding, to the tree data structure, a plurality of branches, wherein each branch of the plurality of branches comprises a synthetic route of the plurality of synthetic routes. 8. The method of claim 1, wherein determining a plurality of synthetic routes associated with the target compound is based on one or more parameters. 9. The method of claim 8, wherein the one or more parameters comprise one or more of available feedstock, available chemical substances, or available equipment. 10. The method of claim 1, wherein determining the one or more optimal synthetic routes from the plurality of synthetic routes is based on one or more parameters. 11. The method of claim 10, wherein the one or more parameters comprise one or more of available feedstock, available chemical substances, available equipment, yield, financial cost, time, reaction conditions, or likelihood of reaction success. 12. The method of claim 1, wherein determining the one or more optimal synthetic routes from the plurality of synthetic routes comprises: determining all compounds that can reach the target in at most a pre-defmed number of steps; and determining a minimal cost synthetic route to the target compound without considering transition telescoping. 13. The method of claim 1, wherein determining the minimal cost route comprises evaluating a cost function. 14. The method of claim 13, wherein the cost function comprises: <img class="EMIRef" id="595953547-imgf000064-0001" /> 15. A method comprising: training, based on a portion of a plurality of known chemical reactions, one or more machine learning classifiers; determining, based on the plurality of known chemical reactions, one or more known chemical reactions that result in a target compound; determining, based on chemical reaction transformations, one or more predicted chemical reactions that result in the target compound, wherein the one or more predicted chemical reactions are predicted as being successful by the one more machine learning classifiers; retrosynthetically determining a plurality of synthetic routes, wherein each synthetic route results in the target compound, wherein at least one synthetic route comprises at least one of the one or more known chemical reactions and at least one of the one or more predicted chemical reactions; and determining, based on a predetermined number of reactions and a cost function, an optimal synthetic route from the plurality of synthetic routes. 16. The method of claim 15, wherein the plurality of known chemical reactions are derived from one or more of, a chemical reaction database, estimated yields, or predicted yields for the one or more chemical reactions. 17. The method of claim 15, wherein training, based on a portion of a plurality of known chemical reactions, one or more machine learning classifiers comprises: receiving a dataset comprising one or more chemical reactions based on one or more chemical transformations, wherein each of the one or more chemical reactions comprises at least one reactant, wherein each reactant is comprised of one or more atoms; for each reactant, classifying the one or more atoms into a category based on a neighborhood atom, a bond order, and/or a number of hydrogen atoms present; for each reactant, determining a vector based on a histogram of categories; determining a training dataset comprised of a) vectors of reactions associated with a specific transformation and b) vectors of reactions associated with the specific transformation but yield a product from a different reaction type; exposing a classifier to a portion of the training dataset to train the classifier; and exposing the trained classifier to another portion of the training dataset to test the trained classifier. 18. The method of claim 17, wherein exposing the trained classifier to another portion of the training dataset to test the trained classifier comprises assessing performance of the trained classifier based on one or more metrics. 19. The method of claim 18, wherein the one or more metrics comprise one or more of accuracy, positive precision, negative precision, positive recall, or negative recall. 20. A system comprising: a computing device, configured to, train, based on a portion of a plurality of known chemical reactions, one or more machine learning classifiers; determine, based on the plurality of known chemical reactions, one or more known chemical reactions that result in a target compound; determine, based on chemical reaction transformations, one or more predicted chemical reactions that result in the target compound, wherein the one or more predicted chemical reactions are predicted as being successful by the one more machine learning classifiers; retrosynthetically determine a plurality of synthetic routes, wherein each synthetic route results in the target compound, wherein at least one synthetic route comprises at least one of the one or more known chemical reactions and at least one of the one or more predicted chemical reactions; and determine, based on a predetermined number of reactions and a cost function, an optimal synthetic route from the plurality of synthetic routes; and a chemical reaction system, in communication with the computing device, configured to, receive the optimal synthetic route, and initiate, based on the optimal synthetic route, one or more chemical reactions.
